CN115310370A - Regional vegetation transpiration prediction method coupling deep learning and physical mechanism - Google Patents
Regional vegetation transpiration prediction method coupling deep learning and physical mechanism Download PDFInfo
- Publication number
- CN115310370A CN115310370A CN202211223868.9A CN202211223868A CN115310370A CN 115310370 A CN115310370 A CN 115310370A CN 202211223868 A CN202211223868 A CN 202211223868A CN 115310370 A CN115310370 A CN 115310370A
- Authority
- CN
- China
- Prior art keywords
- vegetation
- transpiration
- alpha
- transpiration amount
- deep learning
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a regional vegetation transpiration prediction method coupling deep learning and a physical mechanism n And the P-T correction coefficient alpha is replaced by vegetation net radiant flux R nv And vegetation correction factor alpha v (ii) a To accurately estimate the vegetation correction coefficient alpha v Innovatively inputting vegetation, soil and meteorological parameters into a deep neural network to predict alpha v Then the estimated alpha v Importing the measured vegetation transpiration amount and the predicted vegetation transpiration amount to calculate the vegetation transpiration amount, and optimizing alpha in the DNN model by verifying the measured vegetation transpiration amount and the predicted vegetation transpiration amount v Is estimated. Compared with a pure deep learning model, the deep learning model coupled with the physical mechanism can be obviously improvedAnd (3) the simulation precision of the vegetation transpiration amount under the extreme weather condition. The method has important significance in the fields of accurately estimating the water demand of regional vegetation, guiding accurate irrigation of farmlands and the like.
Description
Technical Field
The invention belongs to the technical field of computers, and particularly relates to a regional vegetation transpiration prediction method coupling deep learning and a physical mechanism.
Background
Land transpiration (ET) is the transfer of water from the land surface to the atmosphere, this water exchange usually involving a phase change of the water, from liquid or ice to gas. Plant Transpiration (T) and soil Evaporation (E) exist simultaneously, and plant Transpiration tends to be more complicated than soil Evaporation and water surface Evaporation, and is closely related to the soil environment, the physiological structure of plants, and the atmospheric conditions. At present, the evapotranspiration estimation models aiming at different underlying surfaces comprise a water quantity balance method, a Penman-Monteith formula, a Priestley-Taylor model, an energy balance method and the like. The Priestley-Taylor (P-T) model is a model for calculating the evapotranspiration of the saturated underlying surface of Priestley and Taylor in 1972 under the condition of no advection, and is widely applied due to the fact that input parameters of the model are few.
The Priestley-Taylor (P-T) model is based on the balance of the evaporation water quantity, and is suitable for calculating the model of the evaporation of the saturated underlying surface on the premise of no advection. However, the P-T formula is proposed under the assumption that no advection occurs, which is difficult to satisfy in real situations, and the underlying surface of the flow field always presents a certain non-uniformity, thereby causing the occurrence of advection. In order to eliminate the calculation error caused by basic assumption, a correction coefficient alpha is introduced into the original P-T formula, and the influence of advection is reflected to a certain extent. However, the correction factor α has a large spatial-temporal variability, resulting in a large uncertainty in the calculation result of the evapotranspiration.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provides a regional vegetation transpiration prediction method coupling deep learning and a physical mechanism.
The invention is realized by the following technical scheme:
a regional vegetation transpiration prediction method coupling deep learning and a physical mechanism comprises the following steps:
step 1, acquiring weather, vegetation and soil related parameter information of a covered target area, and obtaining actual measurement vegetation transpiration amount information corresponding to the information;
step 2, improving and constructing a vegetation transpiration model based on the traditional P-T transpiration model, wherein the constructed vegetation transpiration model is represented as follows:
wherein T represents the transpiration amount of vegetation; delta is the slope of the temperature-saturated water vapor pressure curve;𝛾measuring constants for a psychrometer; alpha is alpha v The vegetation correction factor; r nv For net radiant flux of vegetation, R nv Estimated using beer's law;
step 3, constructing prediction alpha v The deep neural network of (a);
the input of the constructed deep neural network is the weather, vegetation and soil related parameters of the target area in the step 1, and the output estimation result is a vegetation correction coefficient alpha v (ii) a Alpha to be estimated v Substituting the vegetation transpiration model constructed in the step 2, calculating to obtain a predicted vegetation transpiration amount, and verifying the predicted vegetation transpiration amount and the actually measured vegetation transpiration amount obtained in the step 1 to optimize the deep neural network;
and 4, predicting the vegetation transpiration amount of the covered target area by utilizing the deep neural network constructed in the step 3 and the vegetation transpiration model constructed in the step 2.
In the above technical solution, in step 1, the acquired weather, vegetation and soil related parameter information includes: air temperature, surface net radiant flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content.
In the technical scheme, in the step 1, a meteorological station system is utilized to observe wind speed, carbon dioxide concentration, carbon dioxide flux, air temperature, surface net radiation flux, relative humidity and saturated water vapor pressure difference; measuring vegetation height and leaf area index by using a plant canopy analyzer; and measuring the soil moisture content of different layers by using a soil moisture meter.
In the technical scheme, in the step 1, the ratio of the vegetation transpiration amount to the total transpiration amount is determined by using a stable hydrogen-oxygen isotope method, and then the vegetation transpiration amount is calculated according to the observed total transpiration amount and is used as the actually measured vegetation transpiration amount.
In the above technical scheme, R nv Estimated using beer's law, see the following equation:
In the technical scheme, in the step 3, a least square method is adopted to calculate an error between the predicted vegetation transpiration amount and the actually measured vegetation transpiration amount obtained in the step 1, if the error is smaller than a specified precision, learning is finished, and the optimal weight and the optimal threshold are output; otherwise, the error signal is propagated reversely along the original connection path and the connection weight and the threshold value of each layer are adjusted step by step until the error is smaller than the designated precision.
In the above technical solution, in step 4, first, acquiring weather, vegetation and soil related parameter information of a target area in a current state, including: air temperature, surface net radiant flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content; then, inputting the parameters into the deep neural network constructed in the step 3, and outputting a vegetation correction coefficient alpha by the deep neural network v (ii) a Finally, the vegetation correction coefficient alpha is corrected v Transpiration model with vegetation
Then calculating the slope delta of the temperature-saturated water vapor pressure curve and the net radiation flux R of the vegetation nv And calculating the vegetation transpiration T of the covered target area under the current state.
The invention has the advantages and beneficial effects that:
the P-T transpiration model is constructed based on an original P-T model, and the original net radiant flux R of the model is used n And the P-T correction coefficient alpha is replaced by vegetation net radiant flux R nv And vegetation correction factor alpha v (ii) a To accurately estimate the vegetation correction coefficient alpha v Creatively combines vegetation, soil and meteorological parametersNumber input Deep Neural Network (DNN) to predict alpha v Then the estimated alpha v Importing the measured vegetation transpiration amount and the predicted vegetation transpiration amount to calculate the vegetation transpiration amount, and optimizing alpha in the DNN model by verifying the measured vegetation transpiration amount and the predicted vegetation transpiration amount v Is estimated. Compared with a pure deep learning model, the deep learning model coupled with the physical mechanism can obviously improve the simulation accuracy of the vegetation transpiration under extreme climatic conditions (extreme drought, heat wave, low leaf area index and cloudy weather). The method has important significance in the fields of accurately estimating the water demand of regional vegetation, guiding accurate irrigation of farmlands and the like.
In addition, it should be noted that, in the prior art, a stable oxyhydrogen isotope method can be used to determine a ratio T/ET (vegetation transpiration amount/total transpiration amount), and then a vorticity-related monitoring device is used to observe the total transpiration amount ET, so as to calculate the actual vegetation transpiration amount T, and this actual measurement method is accurate, but expensive test equipment is used, and a professional experimenter is required to perform an experimental determination in a laboratory (for example, a stable oxyhydrogen isotope method is used, and is required to be performed in a laboratory, and the test equipment is generally tens of thousands of yuan, which is very expensive). Although the method needs to acquire the weather, vegetation and soil related parameter information of the target area, the information can be rapidly detected by using a conventional detection device, and the operation is simple and convenient.
Drawings
Fig. 1 is a flowchart of a regional vegetation transpiration prediction method of the invention that couples deep learning and physical mechanisms.
For a person skilled in the art, without inventive effort, other relevant figures can be derived from the above figures.
Detailed Description
In order to make the technical solution of the present invention better understood, the technical solution of the present invention is further described below with reference to specific examples.
A regional vegetation transpiration prediction method coupling deep learning and physical mechanisms, referring to fig. 1, comprising the following steps:
step 1, aiming at a covered target area, acquiring weather, vegetation and soil related parameter information of the area, wherein the method comprises the following steps: air temperature, surface net radiant flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content. All of the above parameters are intensity or flux parameters.
Specifically, the weather station system can be used for observing information such as wind speed, carbon dioxide concentration, carbon dioxide flux, air temperature, surface net radiant flux, relative humidity, saturated vapor pressure difference and the like; the vegetation height and the leaf area index can be measured by using a plant canopy analyzer; the soil moisture content of different levels can be measured by utilizing the soil moisture meter.
And (3) measuring the ratio of T/ET (vegetation transpiration/total transpiration) by using a stable hydrogen-oxygen isotope method, observing the total transpiration ET by using vorticity-related monitoring equipment, and further calculating to obtain the vegetation transpiration T. The vegetation transpiration amount T is calculated by the measured T/ET ratio and the measured ET, so that the vegetation transpiration amount T can be understood as a measured real value.
And 2, improving and constructing a vegetation transpiration model based on the traditional P-T transpiration model.
The expression of the conventional P-T evapotranspiration model is as follows:
wherein ET is the evapotranspiration amount (mm); Δ is the slope of the temperature-saturated water vapor pressure curve (kPa/° C);𝛾constant (kPa/° C) for wet-dry table measurements; r n For net surface radiant flux (W/m) 2 ) (ii) a G is the soil heat flux (W/m) 2 ) Far less than R n And are generally ignored; alpha is a Priestley-Taylor correction coefficient.
The invention uses the surface net radiation flux R in the traditional P-T evapotranspiration model n And the correction coefficient alpha is replaced by the vegetation net radiant flux R nv And vegetation correction factor alpha v Thereby obtaining a new structureThe vegetation transpiration model is built, and the expression is as follows:
wherein T represents a vegetation transpiration amount (mm); Δ is the slope of the temperature-saturated water vapor pressure curve (kPa/° C);𝛾constant (kPa/° C) for wet-dry table measurements; alpha is alpha v The vegetation correction factor; r nv Net radiant flux (W/m) for vegetation 2 )；R nv Estimated using beer's law, see formula (3) below:
in the formula, R n For net surface radiant flux (W/m) 2 ) (ii) a LAI stands for leaf area index (m) 2 /m 2 ) (ii) a k represents an extinction coefficient; theta s Representing the zenith angle of the sun.
Step 3, constructing prediction alpha v Deep Neural Network (DNN)
The key for calculating the vegetation transpiration amount of the target area by applying the vegetation transpiration model constructed in the step 2 is to accurately determine the vegetation correction coefficient alpha v For the conventional P-T evapotranspiration model, the correction coefficient α is usually set to 1.26 or estimated as a function of the green canopy fraction, solar radiation, however, these methods of determining the correction coefficient have great uncertainty and cannot be applied to α under various vegetation types and meteorological conditions v And (6) estimating. To this end, the present invention innovatively inputs vegetation, soil and weather-related parameters of a target area into a Deep Neural Network (DNN) to predict alpha v And introducing CO into meteorological parameter data set 2 Concentration and flux data to improve the prediction accuracy of the vegetation transpiration amount. Specifically, step 3 includes the following steps.
Step 3.1: constructing a Deep Neural Network (DNN), wherein a DNN algorithm is a multilayer feedforward Neural Network trained according to error reverse propagation and comprises an input layer, an implied layer and an output layer, wherein the input layer (S) comprises N nodes, the implied layer (H) comprises H nodes, the output layer (r) comprises 1 node, N characteristic values of samples are input from the input nodes, forward propagation is carried out, and the output of the nodes of the implied layer is as follows:
wherein w: ( j , k ) Shows the connection weight of the input layer k node to the hidden layer j node, x: ( t , k ) A k-th characteristic value representing the t-th sample of the input, b: ( t , j ) F is a threshold value of a hidden layer, represents an activation function, selects a ReLU function as the activation function, and has the following calculation formula:
the output of the output layer nodes are:
wherein w: ( t , j ) Representing the value of the hidden layer j node to the output node connection weight, b (t) Is the threshold of the output layer.
Step 3.2: the N characteristic values of the input samples of the input nodes of the constructed deep neural network comprise: air temperature, surface net radiation flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content, and the output estimation result is vegetation correction coefficient alpha v α to be estimated v And (3) further substituting the vegetation transpiration model constructed in the step (2) and calculating to obtain the predicted vegetation transpiration amount.
Step 3.3: and (3) verifying the predicted vegetation transpiration amount obtained in the step (3.2) and the actually measured vegetation transpiration amount obtained in the step (1) to optimize the deep neural network. Specifically, in this embodiment, a least square method is used to calculate an error E (θ) between the vegetation transpiration estimated in step 3.2 and the measured vegetation transpiration obtained in step 1, and if the error is smaller than a specified precision, learning is finished, and an optimal weight and a threshold value are output; otherwise, the error signal is reversely propagated along the original connection path and the connection weight and the threshold of each layer are gradually adjusted until the error is smaller than the designated precision.
And 4, predicting the vegetation transpiration amount of the covered target area by utilizing the deep neural network which is constructed in the step 3 and meets the prediction precision requirement and the vegetation transpiration model which is constructed in the step 2.
Firstly, acquiring weather, vegetation and soil related parameter information of a target area in a current state, wherein the acquiring comprises the following steps: air temperature, surface net radiation flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content; then, inputting the parameters into the deep neural network which is constructed in the step 3 and meets the prediction precision requirement, and outputting a vegetation correction coefficient alpha by the deep neural network v (ii) a Finally, the vegetation correction coefficient alpha is corrected v Transpiration model with vegetation
Then calculating the slope delta of the temperature-saturated water vapor pressure curve and the net radiation flux R of the vegetation nv The vegetation transpiration T of the covered target area under the current state can be accurately calculated.
The invention being thus described by way of example, it should be understood that any simple alterations, modifications or other equivalent alterations as would be within the skill of the art without the exercise of inventive faculty, are within the scope of the invention.
Claims (7)
1. A regional vegetation transpiration prediction method coupling deep learning and a physical mechanism is characterized by comprising the following steps:
step 1, acquiring weather, vegetation and soil related parameter information of a covered target area, and obtaining actual measurement vegetation transpiration amount information corresponding to the information;
step 2, improving and constructing a vegetation transpiration model based on the traditional P-T transpiration model, wherein the constructed vegetation transpiration model is represented as follows:
wherein T represents the transpiration amount of vegetation; delta is the slope of the temperature-saturated water vapor pressure curve;𝛾measuring constants for a psychrometer; alpha is alpha v The vegetation correction factor; r nv For net radiant flux of vegetation, R nv Estimated using beer's law;
step 3, constructing prediction alpha v The deep neural network of (a);
the input of the constructed deep neural network is the weather, vegetation and soil related parameters of the target area in the step 1, and the output estimation result is a vegetation correction coefficient alpha v (ii) a Alpha to be estimated v Substituting the vegetation transpiration model constructed in the step 2, calculating to obtain a predicted vegetation transpiration amount, and verifying the predicted vegetation transpiration amount and the actually measured vegetation transpiration amount obtained in the step 1 to optimize the deep neural network;
and 4, predicting the vegetation transpiration amount of the covered target area by utilizing the deep neural network constructed in the step 3 and the vegetation transpiration model constructed in the step 2.
2. The method of claim 1 for regional vegetation transpiration prediction coupled with deep learning and physical mechanisms, wherein: in step 1, the acquired weather, vegetation and soil related parameter information includes: air temperature, surface net radiant flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content.
3. The method of claim 2 for regional vegetation transpiration prediction coupled with deep learning and physical mechanisms, wherein: in the step 1, observing the wind speed, the carbon dioxide concentration and the carbon dioxide flux by using a vorticity related monitoring system; observing air temperature, surface net radiant flux, relative humidity and saturated vapor pressure difference by using a meteorological station system; measuring vegetation height and leaf area index by using a plant canopy analyzer; and measuring the soil moisture content of different layers by using a soil moisture meter.
4. The method of claim 1 for regional vegetation transpiration prediction coupling deep learning and physical mechanisms, wherein: in the step 1, the ratio of the vegetation transpiration amount to the total transpiration amount is determined by using a stable hydrogen-oxygen isotope method, and then the vegetation transpiration amount is calculated according to the observed total transpiration amount and is used as the actually measured vegetation transpiration amount.
5. The method of claim 1 for regional vegetation transpiration prediction coupling deep learning and physical mechanisms, wherein: r is nv Estimated using beer's law, see the following equation:
6. The method of claim 1 for regional vegetation transpiration prediction coupling deep learning and physical mechanisms, wherein: in step 3, calculating the error between the predicted vegetation transpiration amount and the actually measured vegetation transpiration amount obtained in step 1 by adopting a least square method, if the error is smaller than the specified precision, finishing learning, and outputting the optimal weight and threshold value; otherwise, the error signal is propagated reversely along the original connection path and the connection weight and the threshold value of each layer are adjusted step by step until the error is smaller than the designated precision.
7. The method of claim 1 for regional vegetation transpiration prediction coupling deep learning and physical mechanisms, wherein: in step 4, first, the current state is collectedThe weather, vegetation and soil related parameter information of the target area of (1), comprising: air temperature, surface net radiant flux, wind speed, relative humidity, saturated water vapor pressure difference, carbon dioxide concentration, carbon dioxide flux, vegetation height, leaf area index and soil moisture content; then, inputting the parameters into the deep neural network constructed in the step 3, and outputting a vegetation correction coefficient alpha by the deep neural network v (ii) a Finally, the vegetation correction coefficient alpha is corrected v Transpiration model with vegetation
Then calculating the slope delta of the temperature-saturated water vapor pressure curve and the net radiation flux R of the vegetation nv And calculating the vegetation transpiration T of the covered target area under the current state.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211223868.9A CN115310370B (en) | 2022-10-09 | 2022-10-09 | Regional vegetation transpiration prediction method coupled with deep learning and physical mechanism |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202211223868.9A CN115310370B (en) | 2022-10-09 | 2022-10-09 | Regional vegetation transpiration prediction method coupled with deep learning and physical mechanism |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115310370A true CN115310370A (en) | 2022-11-08 |
| CN115310370B CN115310370B (en) | 2023-08-29 |
Family
ID=83866128
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202211223868.9A Active CN115310370B (en) | 2022-10-09 | 2022-10-09 | Regional vegetation transpiration prediction method coupled with deep learning and physical mechanism |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115310370B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116298184A (en) * | 2022-12-07 | 2023-06-23 | 中国科学院南京土壤研究所 | Tracing method for farmland soil evapotranspiration water vapor migration process by utilizing oxyhydrogen isotope |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006345768A (en) * | 2005-06-15 | 2006-12-28 | Kyushu Electric Power Co Inc | Calculating method of evapotranspiration, and watering control method |
| CN111783987A (en) * | 2020-07-14 | 2020-10-16 | 中国水利水电科学研究院 | Farmland reference crop evapotranspiration prediction method based on improved BP neural network |
| CN112163703A (en) * | 2020-09-25 | 2021-01-01 | 中国水利水电科学研究院 | Farmland reference crop evapotranspiration prediction method considering meteorological factor uncertainty |
| CN112529233A (en) * | 2019-09-18 | 2021-03-19 | 中国农业大学 | Method for predicting evapotranspiration amount of lawn reference crops |
| CN114510824A (en) * | 2022-01-11 | 2022-05-17 | 兰州大学 | Construction method of regional scale evapotranspiration model synchronously considering dynamic changes of vegetation canopy and root system |
| CN114970355A (en) * | 2022-06-02 | 2022-08-30 | 石河子大学 | Method for simulating Xinjiang poplar transpiration water consumption based on cross validation and neural network |
-
2022
- 2022-10-09 CN CN202211223868.9A patent/CN115310370B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006345768A (en) * | 2005-06-15 | 2006-12-28 | Kyushu Electric Power Co Inc | Calculating method of evapotranspiration, and watering control method |
| CN112529233A (en) * | 2019-09-18 | 2021-03-19 | 中国农业大学 | Method for predicting evapotranspiration amount of lawn reference crops |
| CN111783987A (en) * | 2020-07-14 | 2020-10-16 | 中国水利水电科学研究院 | Farmland reference crop evapotranspiration prediction method based on improved BP neural network |
| CN112163703A (en) * | 2020-09-25 | 2021-01-01 | 中国水利水电科学研究院 | Farmland reference crop evapotranspiration prediction method considering meteorological factor uncertainty |
| US20220100997A1 (en) * | 2020-09-25 | 2022-03-31 | China Institute Of Water Resources And Hydropower Research | Farmland reference crop evapotranspiration prediction method considering uncertainty of meteorological factors |
| CN114510824A (en) * | 2022-01-11 | 2022-05-17 | 兰州大学 | Construction method of regional scale evapotranspiration model synchronously considering dynamic changes of vegetation canopy and root system |
| CN114970355A (en) * | 2022-06-02 | 2022-08-30 | 石河子大学 | Method for simulating Xinjiang poplar transpiration water consumption based on cross validation and neural network |
Non-Patent Citations (3)
| Title |
|---|
| HAN CHEN ET AL.: "A hybrid deep learning framework with physical process description for simulation of evapotranspiration", 《JOURNAL OF HYDROLOGY》 * |
| 李晗 等: "区域遥感双源蒸散发模型研究进展", 《水资源保护》 * |
| 苏里坦等: "基于改进BP神经网络的干旱区芦苇腾发量预测模型", 《干旱区地理》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116298184A (en) * | 2022-12-07 | 2023-06-23 | 中国科学院南京土壤研究所 | Tracing method for farmland soil evapotranspiration water vapor migration process by utilizing oxyhydrogen isotope |
| CN116298184B (en) * | 2022-12-07 | 2023-11-21 | 中国科学院南京土壤研究所 | Tracing method for farmland soil evapotranspiration water vapor migration process by utilizing oxyhydrogen isotope |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115310370B (en) | 2023-08-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108875254B (en) | One-dimensional variation inversion method of atmospheric temperature wet profile | |
| Singh et al. | Evaluation and generalization of 13 mass‐transfer equations for determining free water evaporation | |
| Crago et al. | A comparison of several evaporation equations | |
| CN108680268B (en) | Bevis model improvement method for area weighted average temperature based on sounding data | |
| CN111737913B (en) | MWHTS clear sky observation bright temperature selection method based on cloud water content inversion | |
| CN112733385B (en) | Water, energy and carbon coupling numerical simulation method and system based on time-varying gain model | |
| CN105588930A (en) | Method for measuring parameters of soil in greenhouse | |
| CN115310370B (en) | Regional vegetation transpiration prediction method coupled with deep learning and physical mechanism | |
| CN105841847B (en) | A kind of method for estimating Surface latent heat fluxes | |
| CN111160680A (en) | Agricultural drought assessment method based on information assimilation and fusion | |
| CN109636677A (en) | Building thermal technique performance estimating method based on model calibration | |
| CN112699621A (en) | Potential evapotranspiration estimation method considering change of underlying surface condition | |
| QIU et al. | Comparison of the three-temperature model and conventional models for estimating transpiration | |
| CN112329334B (en) | MWHTS and MWTS-II fusion inversion sea surface air pressure method based on simulated bright temperature | |
| Kim et al. | Application of the Simple Biosphere Model (SiB2) to a paddy field for a period of growing season in GAME-Tropics | |
| Rutgersson et al. | Net precipitation over the Baltic Sea during present and future climate conditions | |
| CN115238327B (en) | Three-dimensional space methane distribution pattern simulation method based on satellite monitoring | |
| Lee et al. | All‐sky microwave humidity sounder assimilation in the Korean Integrated Model forecast system | |
| CN112529233A (en) | Method for predicting evapotranspiration amount of lawn reference crops | |
| CN114880933A (en) | Atmospheric temperature and humidity profile inversion method and system for non-exploration-site foundation microwave radiometer based on reanalysis data | |
| Andreas | Two experiments on using a scintillometer to infer the surface fluxes of momentum and sensible heat | |
| CN103218532B (en) | Green water flow estimation method and device | |
| CN115270068A (en) | Method for quickly estimating atmospheric moisture delay of offshore inclined path based on buoy platform | |
| CN114037125A (en) | Numerical weather forecasting method based on all-weather observation error covariance matrix | |
| Hu et al. | Research on Reference Crop Evapotranspiration Forecast Based on FOA-GRNN |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |